This thread intrigues me. The excitement it has seemed to generate from everyone is expected. A few easy rules and whola!..... money machine.... Here are my thoughts... 1) The methodology will lose BIG money over time 2) If PM has made money, he states that he doesn't trade all the time. It's his stroke of Random Luck that he has hit the pockets of times when the methodology works 3) Why doesn't anyone test? 4) Why doesn't anyone test? 5) Why doesn't anyone test? 6) There is no magic set of indicators, magic combo of indicators that will pull money out of the market ALL the time 7) Markets are dynamic, about the only thing that stays the same is the spacing in time and price between data points 8) Patterns in indicators come and go 9) Learn from this thread, learn that No_PM has applied strict money management around his methodology 10) Learn that any method at some time will be profitable 11) Learn that it is positive expectancy either through accuracy or solid risk:reward that makes money 12) This method only has solid risk:reward, unfortunately its accuracy is way too low to achieve profits 13) Its a negative expectancy 14) The very nature of releasing an "edge" alters the edge's balance in the market place. The more people know, the less effective it becomes. No way around this. 15) If NO_PM really had an edge, why in hell would you release it 16) Making money in the market requires a positive expectancy, experience, and emoitional control 17) Unfortunately, the industry is moving towards systematic trading.... strategyrunner.com.. so more and more people are looking for the next system 18) Go read Marketsurfer. His buy on the crisis play is what people should be trying to learn how to do.... 19) Read minyanville.com... Harrison while being bearish has the thought process of a trader... 20) Wish it was as easy as some companies and authors put out that buying a system makes money automatically Good luck... BTW, the system has been tested over 4 years of data, long and short side... its a consistent loser.... sorry guys but I had to have my guys test it.... it just seemed too easy....
WOW, Thanks for being so candid. I like your style, right between the eyes, saves time and energy. Get right down to business. Interesting comments. BTW, very nice post here Offshore... http://www.elitetrader.com/vb/showthread.php?s=&postid=251618#post251618
Thanks... the fatal flaw in the whole system is not statistically testing how many times the 5 minute market actually trades between keltner bands on a 2.5 multiplier...... The system is asking if the market bounces from one side of the band to the other side when divergence occurs in a trending market.... First thing to test is to see if the market does indeed bounce between a 2.5 atr band when the ADX hits 30.... if one can come up with a statistic edge here you might have a good system.... If you can come up with say a 40% probability you get a trading range after the ADX hits 30, then you must be able to detect the turns in the market... then you must set up advantageous risk:reward... then you must be able to define a volatility stop so you don't get shaken out....but the whole system is based upon the ping pong effect.... and that is the flaw... there is no statistical edge when ADX>30 and the market turning into PING PONG.... Food for thought for everyone! Boy, saturday night in front of a computer... hmm... do i have any friends?? Hahahaha I like quiet summer nights to think...
Can you publish the hard data on your tests here? I'm sure that it would interest all of us, including no pm. If your guys have tested it over that long of a period, they've either automated it, or have collected records on checking every potential setup manually. Publishing the info would go a long way towards our education. Thanks.
I'm having trouble exporting the results on this computer but here are the main stats.... Tested over the SP Continuous data for the last two years.... total net profit before commissions ($42,125.00) Total number of trades : 222 Winning trades:54 Losing Trades: 168 percent profitable: 24.32% max cons winning trades: 8 max cons losing trades: 17 Avg Winner: 1811.57 Avg Loser (833.04) Max Drawdown 54,050.00 Max Runup $11,075.00 Percent in market 1.94% These are the main stats I picked from the summary... When I can get the damn thing to export I'll upload the whole summary... The Performance graph is straight down....
OffShoreTrader, If you programmed this in Tradestation, would you mind posting the code (or PMing it to me). Thanks, Richard
inputs: Price( Close ), KeltnerLength( 20 ), Factor( 2.5 ), fastlength(3), slowlength(10), adxlength(14),risk(3); variables: Avg( 0 ), Shift( 0 ), LowerBand( 0 ), UpperBand( 0 ),valadx(0), condition1(false), condition2(false), condition3(false),condition4(false),stopprice(0),exitprice1(0); condition1 = false; condition2 = false; condition3 = false; condition4 = false; {Determine Keltner Values} Avg = AverageFC( Price, KeltnerLength ) ; Shift = Factor * AvgTrueRange( KeltnerLength ) ; UpperBand = Avg + Shift ; LowerBand = Avg - Shift ; {Determine ADX Value} ValADX = ADX(ADXLength); {Determine setup} if BearishDivergence(High, chaikinosc(volume, FastLength, SlowLength), 1, 20) = 1 then condition1 = true; if valadx > 30 then condition2=true; if high < high[1] then condition3=true; if high > highest(upperband,10) then condition4 = true; {Short Order} if condition1 and condition2 and condition3 and condition4 and time < 1100 and marketposition = 0 then sellshort ("Short Div") next bar at open; {Manage trade} if marketposition = -1 then begin stopprice=entryprice+((upperband-lowerband)/Risk); exitprice1=lowerband; buytocover ("Short Stop") next bar at stopprice on a stop; buytocover ("Profit") next bar at exitprice1 on a limit; end; This is the short side of the system...
Perhaps you can tell us which testing platform you are using. We have experts here who are experienced with a variety of platforms. If you clue us in on how this testing was done, we can help you with exporting the data. And as rickty said, we'd all like to see the code used for the testing. Thanks again. Mark Edit - looks like you beat me to it!